Bhopa, India – Case Study

Journal ofRisk and Uncertainty. 12:257-269 (1996) © 1996 Kluwer Academic Publishers

Union Carbide’s Bhopal Incident: A Retrospective

MICHAEL J. FISCHER Marsh & McLennan, Inc.. 1166 Avenue of the Americas, New York, NY 10036-2774

Abstract

This article examines the actual events ofthe Bhopal gas disaster as a prelude to considering the social treatment of catastrophic risks of this variety. In the context of the conference. Bhopal is held out to be symbolic of man-made catastrophes having sudden impact and is therefore the contrasting episode to an examination of the symbolic episode for gradual man-made catastrophes, asbestos liability. Bhopal is then connected to the cir- cumstances which created a shortage of commercial catastrophe liability insurance in the 1980s. Its influence on the restructuring of the market is then discussed. A commentary on the buying habits of large industrial corporations is also included.

Key words: MIC. Excess Liability, Bhopal Act, ACE, XL

It has been nearly ten years since “Bhopal” entered our lexicon and came to symbolize a worst case industrial catastrophe. It ranks with Exxon’s Valdez as an event which came to embody many of our worst fears about the catastrophic risks we incur every day.

Nevertheless, Bhopal is also perhaps one ofthe least understood episodes in the history of industrial accidents. Its occurrence involved a combination of many factors, some of which were unique to the Bhopal plant. The primary purpose of this article, therefore, is to provide some insight into what actually happened, and, in so doing, to provide a factual basis from which the factors contributing to the event can be properly evaluated.

I will then link the Bhopal incident with developments in the market for commercial liability insurance in the mid-1980s. This market suffered a severe contraction and was significantly restructured after Bhopal. My secondary purpose is to demonstrate the im- pact which Bhopal had on the changes the excess liability market underwent and, in so doing, to describe the current market for catastrophic liability risks.

As a final point, I offer some views conceming the present demand for high levels of excess liability insurance and what this may mean in terms of social poiicy towards catastrophic risk and how this contrasts with other catastrophe exposures treated in this conference, specifically asbestos and hurricanes.

1. The event

Early on December 3, 1984, a chemical reaction in storage tank 610 at the Union Carbide India pesticide plant in Bhopal, India caused the tank’s safety relief valve to open. Inside

 

 

2 5 8 MICHAEL J. FISCHER

tank 610 were some 41 metric tons (equating to 90,000 pounds) of methyl isocyanate (MIC), a toxic intermediate chemical used in the processing of Carbide’s SEVIN-brand carbaryl pesticide.

A vigorous exothermic reaction occurred when a large quantity of water (believed to be between 1,000 and 2,000 pounds) was introduced into the tank and mixed with the 90,000 pounds of stored MIC. A concurrent exothermic trimerization of MIC (catalyzed by iron resulting from corrosion of the stainless steel tank walls under the high temperatures, produced by the initial reaction) kept the safety relief valve open for approximately two hours. During that time, more than 50,000 pounds of MIC in vapor and liquid form were discharged into the atmosphere (UCC Team Report 1985). The magnitude of this release overwhelmed the vent gas scrubber and fire hoses employed to contain minor leaks, allowing a toxic cloud of MIC gas to escape over the plant gates into the area surrounding the plant. This surrounding area was a densely populated and impoverished district where thousands of people lived in makeshift huts.

This overview of the incident was pieced together by investigators from Union Carbide Corporation, which owned 50.9% of the company that operated the plant (Union Carbide India Limited), and engineering consultants Arthur D. Little Inc. Together they conducted a painstakingly detailed investigation of the events that led to the deaths of between 1,500 and 4,000 people and injured thousands more in what has been called the world’s worst industrial accident.

The Bhopal tragedy has many dimensions. It is foremost a human story, given both the terrible loss of life and the human element that proved to be the catalyst for the disaster. But it is also a political story, a detective story, and a story about the consequences which the disaster had on one of the largest chemical companies in the world. Any one of these facets can be explored in great depth, but that is not the focus of this article. There is, however, an element involving the investigation of the disaster that resembles a good detective story and is critical to understanding the event in the context of social policy relating to catastrophic risks.

MIC is a key ingredient in the manufacture of agricultural pesticides, which incidently played a significant role in India’s “green revolution.” Estimates are that pesticides saved about 10% of the annual food crop in India at the time of the Bhopal incident—enough to feed over 70 million people (UCC Annual Report, 1984).

LL MIC chemistry

The raw materials used to make MIC are monomethylamine (MMA) and phosgene. In Bhopal, phosgene was produced on-site by reacting chlorine and carbon monoxide (CO). CO was produced at an adjacent production facility within the plant. MMA and chlorine were brought in by tank truck from other parts of India, stored in tanks, and used as needed. Chloroform was used as a solvent throughout the process. A vapor phase reaction system converted the phosgene and MMA to methyl carbamoyl chloride (MMC) and hydrogen chloride (HCL).

 

 

UNION CARBIDE’S BHOPAL INCIDENT: A RETROSPECTIVE 259

The reaction products were then quenched with chloroform and fed to a phosgene stripping still where the unreacted phosgene was removed and recycled. The product was then fed to a pyrolyzer to separate the MIC from the hydrogen chloride.

The MIC was then refined in a still where it was purified through separation from chloroform, MMC, and other residues. The refined MIC was then stored in two of three 15,000-gallon stainless-steel tanks. The third tank was used for emergency storage of MIC and for temporarily holding off-specification MIC prior to reprocessing. The contents of tanks were circulated through heat exchangers cooled by a 30-ton refrigeration system which maintained the MIC at a temperature of about 0° C (UCC Team Report, 1985).

Extensive precautions were taken throughout the chemical process to avoid contami- nation of the stored MIC. Of particular concern was the risk of water contamination, since MIC reacts with water. Tanks and lines were dried with high-purity nitrogen, and the refrigerant used to cool stored MIC was nonaqueous to eliminate the possibility of water contamination. In a nutshell, it was well known among plant operators that water and MIC should not be mixed (Kalelkar, 1988).

L2. The actual cause

This brings us to the central story of what really caused the Bhopal disaster. It is not possible to consider Bhopal in the context of managing catastrophic risks without a clear understanding of what actually happened.

Original press reports seized on a theory embraced by India’s Central Bureau of In- vestigation (CBI), that an improper water-washing of the relief-valve vent header in the MIC manufacturing unit had caused water to enter the tank, initiating the reaction. It was believed that an MIC operator had failed to insert a slip-bind, as called for by plant operating procedures, to prevent water from backing up into the process vent line and contaminating the system of pipes, tubes, and valves throughout the unit. This early theory of the disaster’s cause was widely reported and accepted. There are many reasons why this happened, not least the phenomenon of reporters drawing hasty conclusions to “scoop” the story. But the theory gained popular appeal in part due to its plausibility^—that a minor inadvertence, forgetting to insert a slip-blind, precipitated a terrible tragedy. The theory also fit with the view suggested by some that perhaps practices at the Bhopal plant—and at other plants throughout the developing world—were not up to the high standards of plants operating inside the United States. Proponents of this view suggested that some countries were not able to absorb the technology brought in by sophisticated multinational firms.

Although Union Carbide investigators (including A. D. Little) arrived at the site within days of the event to provide assistance and conduct an investigation, they found that the plant had been sealed and placed under control of the CBI (Kalelkar, 1988).

Prior to beginning their investigation, both Carbide and the CBI determined that 20 tons of MIC that remained in a second tank (tank 611) had to be disposed of and that the best way to accomplish this was to process it into pesticide (Kalelkar, 1988). The processing of the MIC was done jointly by Union Carbide India Limited and the Goverrunent of

 

 

260 MICHAEL J. FISCHER

India, with the assistance of the Carbide technical team, in an activity labeled “Operation Faith” by the Indian government (Kalelkar, 1988).

After this process was successfully completed, the Carbide team was permitted to begin its investigation, but encountered numerous obstacles placed by India’s CBI, which had taken control of the plant and its records. CBI also placed Warren Anderson, Carbide’s chairman at the time (who went to Bhopal to offer assistance and relief immediately afrer the incident), under arrest, along with seven other UCIL officers and employees (Kalelkar, 1988). This activity further impeded the investigation, as plant employees, fearful of criminal sanctions, became less forthcoming.

Despite all of these handicaps. Carbide investigators ultimately determined the precise cause of the disaster. Through extensive interviews and experimentation, investigators determined that the following sequence of events had taken place (UCC Team Report 1985).

• At 10:20 PM. on the night of the incident, the pressure in tank 610 was at 2 psig. (This is significant because no water could have entered the tank prior to that point; other- wise, a reaction would have begun, and the resulting pressure rise would have been noticed).

• At 10:45 PM., a shift change occurred. The unit was shut down, and the shift change took at least a half hour to be accomplished.

• Investigators believe it was this point—during the shift change—that a disgruntled employee entered the storage area and hooked up a rubber water hose to tank 610, with the intention of contaminating and spoiling the tank’s contents.

• At 11:00 PM., the control room operator noticed that the pressure in tank 610 was at 10 psig. This pressure was not thought to be unusual, because the tank was normally operated at a pressure between 2 and 25 psig.

• At 11:00 PM., the field operator reported a small MIC leak near the vent gas scrubber, but operating personnel did not discover the source of the leak. Since minor leaks occurred from time to time, it would not have been unusual for the field operator to conclude that there wasn’t major problem.

• The reaction of the water and MIC initiated the formation of carbon dioxide, which, together with MIC vapors, was carried through the header system and released into the atmosphere out of the vent gas scrubber.

• By 11:30 PM., workers downwind of the unit sensed MIC vapors and reported the leak to the supervisor. Workers began the search for the source.

• About midnight, field operators found what they believed to be the source—a section of open piping on the second level of the unit—and fixed a fire hose to spray in the direction of the leak, believing that they had successfully contained the leak. They didn’t realize that this relatively minor release was just the tip of the iceberg.

• Shortly after midnight, several MIC operators saw the pressure rise in tank 610 to a point where it was off the scale (over 55 psig).

• Control room operators ran to the tank, from which they heard rumblings, as well as a screeching noise from the relief valve, and also felt heat radiating. At that time, they also discovered the water hose connection to the tank, and realized that they had a grave situation on their hands. They activated the vent gas scrubber (which had been on

 

 

UNION CARBIDE’S BHOPAL INCIDENT: A RETROSPECTIVE 2 6 1

Standby mode) and attempted to remove the water from the tank—actions that were futile at this point. (Not knowing whether or not their attempts to remove the water had exacerbated the situation, those involved later participated in a “cover-up” of the actual events, giving rise to the original water-washing theory.)

• It was then, around 12:15 A.M. on Monday, December 3rd, that the major release occurred.

This conclusion as to the precise sequence of events is critical to understanding the Bhopal story and its effect on Union Carbide, a company that had an excellent record for safety. The fact that this tragedy was caused by an act of employee sabotage provided a counterpoint to the often overblown rhetoric that faulted Carbide for transferring sophis- ticated technology to a developing country lacking the infrastructure to absorb it. Carbide utilized this discovery to help restore some balance to the public debate and to its standing in the world business community, especially in the financial and insurance community.

2. The settlement

Following the realization of the full extent of the disaster. Carbide, which had provided immediate aid and relief to the Bhopal victims, pursued a quick and fair settlement of all claims arising from the disaster. However, negotiations with the Union of India, which had passed the Bhopal Act, making the govemment the sole representative of the victims, were repeatedly stymied by political dynamics in India. In 1989, unable to reach an accord directly with the government and faced with the prospect of endless litigation and a demand for $3 billion in damages (a sum exceeding Carbide’s entire net worth). Carbide agreed to accept a settlement of the Bhopal loss as ordered by the Supreme Court of India. The Court’s order, dated February 14, 1989 read: “Having given our careful consideration for these several days to the facts and circumstances of the case placed by the parties in these proceedings, the mass of data placed before us, the material relating to the proceed- ings in the Courts in the United States of America, the offers and counter-offers made between the parties at difFerent stages during the various proceedings, as well as the complex issues of law and fact raised before us and the submissions made thereon, and in particular the enormity of human suffering occasioned by the Bhopal Gas disaster and the pressing urgency to provide immediate and substantial relief to victims of the disaster, we are of the opinion that the case is pre-eminently fit for an overall settlement between the parties…” The price tag was $470 million, exclusive of legal fees, $420 million paid by Union Carbide Chemicals & Plastics, with Union Carbide India Limited paying the Rupee equivalent of $45 million. Carbide was credited with an additional $5-million amount, previously paid to the Red Cross at the suggestion of U S . Judge John F. Keenan. The Union of India accepted the settlement on behalf of all victims of the tragedy under the Bhopal Act, which the Supreme Court of India later upheld. Carbide Chairman Robert D. Kennedy (who succeeded W.A.) said in a February 14, 1989 news release, “we are pleased that this will now provide for the care and rehabilitation of the victims and their families and a fair resolution of all issues.”

 

 

2 6 2 MICHAEL J. FISCHER

3. Carbide coverage

The Bhopal loss far exceeded the catastrophe liability coverage which Carbide had in place at the time of the incident. The $200 million “umbrella” program that Carbide maintained in 1984 was fairly typical for a Fortune 100 company at that time. While some companies purchased higher limits (and some had less), $200 million was a seemingly fair and reasonable amount of coverage to carry, even for a large chemical manufacturer. But viewed against the magnitude of the Bhopal loss, it now seems almost preposterous to suggest $200 million of coverage was “fair and reasonable.” Certainly, in hindsight, it can be said that Carbide would have been better off if it had purchased every nickel of coverage which it could get, regardless of price. But even given the magnitude of the risks exposed by the tragedy, it must be pointed out that prior to the loss Carbide had a sterling loss record—they had never experienced any loss near even a small fraction of the Bhopal loss. Based on a long-standing record of process safety and no historical precedent in the way of catastrophic loss, I think it is fair to say Carbide had reason to believe that they were well protected.

4. The 1980s market

1984 was a “buyers” market for insurance in the U.S., as casualty premiums in particular dropped to exceedingly low levels under the “cash flow” underwriting philosophy that was then prevalent. Buyers of excess liability coverage tended to focus largely on price; security of the carders was a secondary issue. The market was soft, capacity was ample, and coverage terms were generally not a problem.

However, under the surface of the market, big problems were lurking. Underwriters were in fact losing a great deal of money; only investment income kept them in the black. The softness of the market masked the fiscal problems ahead. The market’s softness had persisted, due in large part to the questionable fundamentals of the catastrophe liability market in the early 1980s.

As Mr. Sinfield points out in his article on asbestos, the “second wave” of asbestos bodily injury claims were being triggered during this period and were beginning to be felt by the industry. Together with the first wave of environmental claims, property/casualty insurer results were beginning to be negatively impacted under the cumulative weight of these “long-tail” liability losses at the time of the Bhopal disaster. In 1984, the consoli- dated U.S. property/casualty insurance industry generated a net operating loss of $3.8 billion, as underwriting losses of $21.5 billion outstripped investment income of $17.7 billion. 1985 was an even worse year, as U.S. underwriters lost $5.1 billion, as losses grew to nearly $25 billion. But conditions before the close of the 1984 underwriting year did not, of course, reflect these dismal results. Thus, 1984 would prove to be the last year of the soft underwriting cycle.

Union Carbide’s excess liability insurance program in 1984 reflected this last round of softness and was fairly representative of the breed for major industrial risks, not only in terms of limit purchased but also in terms of structure.

 

 

UNION CARBIDE’S BHOPAL INCIDENT: A RETROSPECTIVE 2 6 3

Carbide’s coverage was organized in eight layers syndicated among 30 insurers, in- cluding syndicates at Lloyd’s of London. Most carriers wrote policies of $5-10 million, scattered throughout the layered “tower” of coverage. The pricing was very low—the rates above the first two layers of coverage (above $23.5 million) averaged just $2,000 for each $ 1,000,000 of limits provided (less than what an A-rated company pays for a letter of credit today), and the scope of coverage was worldwide.

The insurers who participated in Carbide’s and similar programs occupied an arcane comer of the property/casualty marketplace. Their premium writings were not a major source of revenue, and the markets relied heavily on reinsurance to provide the capacity needed. The underwriting process chiefly focused on an insured’s loss record, influenced somewhat by an element of “class of business” underwriting. Certainly, large chemical companies had a slightly more difficult time piecing together their programs, and paid somewhat higher prices than “plain vanilla” risks in the 1984 market. They were also more likely to have had to discuss their exposures more completely, particularly if a specific chemical was making headlines. But underwriters were generally more concerned with the possibility of having to “drop down” into a low (closer to loss) layer if underlying policies’ limits were exhausted by moderately large losses than they were of the “big hit.” Underwriters also tended to look mostly at U.S. exposures; overseas exposures were a throw-in in terms of the coverage provided.

5. Complexity of risk

In point of fact, chemical companies present complex risks, and the process technology involved is beyond the grasp of the average insurance professional. In the early 1980s the insurance market responded to this complexity with a crude instrument—the open or closed window. Carriers that concluded that the risks were too difficult to understand would simply not underwrite a chemical company at any price or on any terms. Those that did not prohibit the consideration of a chemical risk underwrote that risk in much the same way that any other risk was reviewed. In addition, many of the companies that provided capacity did so with little exposure, because they reinsured most of the risk.

6. lYansition

Prior to 1985, the chemical industry, and the Fortune 500 companies that comprised a large segment of the demand for excess liability coverage, had no problem obtaining capacity. But, in 1985, the market all but evaporated for high levels of coverage. There are a few key reasons why this happened.

Carriers who had provided this coverage relied extensively on reinsurance. Once rein- surers stopped accepting risks, the capacity dried up, and the market suddenly contracted. But something more insidious had a major role in the market contraction; that was the “defective” policy under which coverage had been provided (Clements, 1994). This de- fective policy is also central to the asbestos story.

 

 

264 MICHAEL J. FISCHER

The policy in question is called an occurrence policy* Essentially this insurance policy can be triggered at any time once a claim is reported alleging that something happened when the policy was in force. This meant that policies written years ago to cover basic accident risks were open to new claims under novel and often expansive theories of liability. The impact of this flaw was brought home in the Keene decision, which made it possible to make claims against policies written over an extended period of time to cover big, murky liability issues such as asbestos or pollution. By making claims against poli- cies written over many years, insureds could “stack” even modest single-year limits into a very large coverage pool. Once underwriters realized what was happening to the policies which they were routinely offering, and given the fact that the coverage was not a sig- nificant source of revenue, underwriters beat a hasty retreat from the excess liability market.

Nevertheless, there is still no question in my mind that Bhopal was a contributing factor in the demise of the excess liability market in 1985 and 1986.

Prior to Bhopal, it was difficult to even imagine a loss of such immensity. One need only look at the modest sums which underwriters were willing to accept to provide coverage in high excess layers to see that those writing the policies never believed that a Bhopal-like disaster could happen. Bhopal crystallized in everyone’s mind the reality of catastrophic potential. Clearly, the fragmented and undercapitalized excess liability spe- cialty carriers were ill-equipped to effectively underwrite complex, highly technological risks. Bhopal made it acutely clear that it was not sufficient to focus on loss records, because even companies with the highest safety record, such as Union Carbide, had the potential for a catastrophic loss.

In contrast to the changes in the tort system and other evolving social factors that are at the root of the asbestos catastrophe (as well as liability from pollution cleanup or more recently silicone breast implants), Bhopal represented a sudden, unexpected event defin- able at a precise moment in time. This type of loss, the “big hit,” has always been, and continues to be, something the private insurance market is prepared to cover. This is the fundamental difference between asbestos and Bhopal when considering social policy towards catastrophic risks. Whereas it can be argued that the insurance market never collected a premium to cover aggregated asbestos and pollution claims in the billions of dollars, underwriters theoretically price for the single major event.

The issue for the private market for risk transfer as it relates to Bhopal is therefore largely a structural one having two components:

1. The way in which the basic agreement between insured and insurer is struck (Clem- ents, 1994), and

2. The way in which the capital needed to absorb catastrophic risks is organized.

* r \ e elected to describe the policy as “defective.” as opposed to focusing on the litigioiisness in our society and the civil justice system which has wrecked haxoc with policy wordings. It certainly can be argued that the policies were not the problem. Howe\er. under the circumstances, it transpired that the policies were in fact defective, since they did not withstand the changes that tested the policy’s language.

 

 

UNION CARBIDES BHOPAL INCIDENT: A RETROSPECTIVE 2 6 5

In 1985, in response to the needs of clients who could no longer obtain levels of excess liability coverage higher than $100 million, Marsh & McLennan joined with J. P. Morgan to form A.C.E. Insurance Company, Ltd.

This new insurer, originally established in the Cayman Islands due to its favorable regulatory climate (not to mention the difficulty of organizing a new U. S. insurer and obtaining the approval of 50 state regulators), was initially capitalized with $285 million contributed by 34 large U.S. corporations (Redmond, 1992). As a policyholder-owned facility, A.C.E. was designed to provide up to $100 million of coverage above the first $100 million of risk (in other words, to cover losses that ranged between $100 million and $200 million). Later, the $100-million threshold served as a minimum point at which A.C.E. would begin its policy, and the amount of coverage offered expanded to $200 million.

This new facility represented a fundamental shift in the structure of the excess liability market. Among other things, it offered:

1. A fairer deal between insured and insurer, with the introduction of something called an occurrence-reported policy. Essentially this new policy eliminated the “stacking” of limits problem that plagued the so-called “occurrence” underwriters. The new policy also contained a “bad actor” exclusion, which removed from coverage losses involving, for example, the health effects of tobacco, the IUD, and any products containing asbestos. Finally, the policy put a specific time limit on pollution incidents to keep gradual pollution claims out of the policies.

2. A sizable capital infusion concentrated in one market. 3. A new underwriting discipline fostered by the concentration of many large risks

reviewed by the same underwriting team, quoting with the same terms and conditions. 4. Net capacity not subject to reinsurance.

In brief, all the flaws that had converged to create the liability crisis of 1985 were factored into the new facility. This gave A.C.E. a great opportunity to succeed in a market that had just collapsed.

Interestingly, Union Carbide was not one of the 34 original sponsoring organizations. Although Carbide offered to invest in the new entity, the A.C.E. selection committee declined to accept them as a sponsor. This declination was a direct result of both Bhopal and a much publicized incident at Carbide’s Institute, a West Virginia “Bhopal sister” operation, in March 1985. This incident, while not resulting in a serious loss, fueled the perception at the time that Carbide was an unfavorable risk.

Raising the capital commitments necessary to launch A.C.E. was not an easy task, largely because of the newness of the concept and the fact that most U.S. industrials were not comfortable getting into the insurance business. Most of the 34 original sponsoring companies had the heaviest exposures, especially the petrochemical and pharmaceutical sectors, and therefore were in most need of the coverage.

Dow, Du Pont, Shell Oil, and Tenneco were among the petrochemical sponsors. Car- bide’s absence from this group is an ironic footnote to the A.C.E. story—a silent reminder of the effect Bhopal had on the catastrophe liability market—on the one hand, giving

 

 

2 6 6 MICHAEL J. FISCHER

added impetus to the drive to start a new insurance facility and, on the other hand, keeping Carbide out of the original equation, so that potential investors wouldn’t be scared away.

Though not accepted as a sponsor. Carbide later applied for coverage and became an A.C.E. policyholder in early 1986. By then, the stigma of Bhopal had begun to fade. The A.C.E. policy provided Carbide with a level of protection which it had gone without since March of 1985.

A.C.E.’s launch did not solve all the problems of buyers of excess liability coverage. Whereas the market for coverage above $100 million collapsed in 1985, the market below $100 million imploded on January 1, 1986, with the expiration of many reinsurance treaties that were not renewed. Just as large companies were once again able to secure the high excess limits which they sought, the bottom fell out: companies now had severe difficulty insuring their risks in the first $100 million of potential loss. Many of the insurers who continued to provide coverage in the first $100 million of risk in 1985 ran into financial trouble in 1986 and ceased writing business. Some went into receivership and were later liquidated, such as Transit Casualty, Integrity Insurance, and Mission Insurance. Most simply withdrew from the market. This predicament gave rise to the “son of A.C.E.” concept.

Building on the success of A.C.E., Marsh & McLennan and J.P Morgan formed a second offshore facility, X.L. Insurance Company, Ltd., in May of 1986. X.L.’s mission was to write the layer of coverage immediately below A.C.E., with a maximum limit of $75 million and a minimum attachment of $25 million.

X.L.’s birth furthered the restructuring of the market for excess coverage that began with A.C.E. It reflected many of the same coverage features and a similar underwriting approach. This approach focused heavily on classes of business and the hazards associated with each class. Risks were assigned hazard ratings and measured against other risks within their class. This underwriting regimen represented a dramatic shift in practices, and was made possible by the large concentration of risks in one, well capitalized market.

This time Carbide was one of the sponsors. Carbide had just successfully defeated a hostile take over by GAF and was moving forward with a restructuring plan; the victory over GAF further repaired Carbide’s reputation in the wake of Bhopal. Carbide also had been working with other chemical firms to form their own industry insurance facility in early 1986. However, when the drive to form X.L. began to build momentum, the chemi- cal companies’ brainchild yielded to the X.L. concept, adding the critical mass necessary for X.L.’s successful launch.

6.7. Dramatic impact

The formation of A.C.E., and later X.L., has had a dramatic impact on the excess liability market. The concept of forming new insurance companies with a clean slate and substan- tial infusions of fresh capital has since given rise to the formation of many other new facilities. Most recently, the crisis in the property catastrophe market brought on by the

 

 

UNION CARBIDE’S BHOPAL INCIDENT: A RETROSPECTIVE 2 6 7

unprecedented frequency and severity of natural disasters has led to the formation of several new Bermuda-based reinsurers. Approximately $4 billion of new capital has flowed to Bermuda property/casualty insurers in 1993-1994 (ACE Annual Report, 1993).

7. Lessons of Bhopal

Bhopal shook both Carbide and the insurance industry to their very cores, and the rever- berations were felt for many years. What is interesting about the unfolding of the Bhopal story is its incredible timing, occurring as it did on the eve of an emerging crisis in the market for commercial liability insurance. Bhopal and the liability insurance crisis flowed together like a torrent, making it difficult to separate the two issues. The magnitude of the changes in the excess catastrophe market since 1986 is so overwhelming that it is difficult to find the legacy of Bhopal contained within those changes.

Under close scrutiny, one can find the unmistakable imprint of Bhopal. Among other things, the Bhopal tragedy:

• Precipitated, or at the very least, accentuated the collapse of the excess liability market in 1985 and 1986.

• Gave impetus to the formation of new insurers which drew in fresh capital to the industry unencumbered by the asbestos and pollution legacies.

• Ultimately transformed the Union Carbide Corporation. • Made chemical companies and the E.P.A. reassess the risks posed by the chemical

industry. • Focused debate on chemical company disclosure and a community’s right to know, with

respect to worst case scenarios at chemical plants in areas close to where people live.

7.1. Frequency versus severity

Bhopal, let us hope, is a one-in-a-million incident that will not happen again. Our society and our standard of living are in many ways tied to the products that the chemical industry provides. From plastics to pesticides, we’ve come to depend on chemicals for our every- day way of life.

The manufacture of these chemicals often entails a considerable amount of risk, and we can all be thankful that the industry has such a strong record for safety and a dynamic safety culture. There is today a well capitalized private insurance market available to absorb the risks posed by the chemical industry, and industry generally. Today it is possible to obtain coverage of $500 million, perhaps somewhat more, at reasonable prices and terms. While much of the available capacity is centered in Bermuda, European insurers and Lloyd’s of London can also provide high amounts of coverage.

 

 

268 MICHAEL J. FISCHER

The fundamental reason that this capacity is available is the relative infrequency of truly catastrophic losses in the liability area. Whereas the frequency of large hurricane and earthquake losses has seen a dramatic rise, only a relatively modest rise is detectable when it comes to “moment in time” losses near the magnitude of a Bhopal.

7.2. Implications for social policy

An adequate mechanism for just compensation of victims in the event of accidental loss is already in place. Large industrial concerns have the opportunity to obtain substantial amounts of liability protection. It is not uncommon for high hazard risks to carry $300 million, or more, of coverage. The problem seems to be that many companies continue to spend their premium budgets on coverage in relatively low layers, say between $10 million and $50 million. It has taken some time for many companies to bring their level of risk retention up to the $10-$25 million range, and many are not comfortable with the idea of a higher level of risk retention. In fact, many large firms with the financial strength to absorb high levels of risk maintain retentions well below $10 million. Oftentimes, this is because the commercial market in many respects still prefers to write moderately high limits in low layers, which keeps premiums in this area attractive. But buyers also must recognize that, above a company’s last (or highest) dollar of coverage, companies are also retaining risk.

Many companies today pursue this strategy of buying coverage in the first $50 million and self-assuming anything above their maximum limit, usually between $200 and $300 million. This indicates that while an insured can imagine a $50-million loss, they have difficulty imagining a $500 million loss. Otherwise, they’d buy to protect against the $500-million loss and self-assume the first $50 million. A survey of A.C.E. policyholders confirms this.”* Since A.C.E. always writes at least a portion of its coverage as the top layer and writes more than half of the Fortune 250 companies, they provide a good picture of buying strategies (see chart).

A\erage limits purchased by A.C.E. insureds

I Oil class S374 Million 2. Pharmaceutical class S324 Million 3. Transportation class S3 I I Million 4. Chemical class S299 Million 5. Construction class S285 Million 6. Utilities class S273 Million 7. Other classes S256 Million

Some risk managers have concluded that they should insure against the truly catastrophic event and buy quality coverage as high as it goes, as long as it is priced reasonably. But there are some who simply won’t buy more, because they don’t believe that a monumental disaster can happen to them. Most will be correct in this assessment, but a few may end up misjudging their maximum exposure and will be underinsured. The financial conse-

 

 

UNION CARBIDES BHOP.AL INCIDENT: A RETROSPECTIVE 269

quences to a corporation of a loss that is underinsured by several hundred million dollars can be severe—even for very large and profitable firms. The effect of a major charge against earnings may be felt by share owners, pensioners, employees, and communities. In this regard, the stabilizing influence of catastrophe liability insurance is an important social policy consideration.

Presently, the cost to obtain this insurance is not excessive, and, even in a zero-sum world, the savings produced by assuming a higher retention can often fund a sizeable increase in a company’s liability protection. Furthermore, the economics of high levels of catastrophe liability insurance are likely to be more favorable over the long term than the economics of insurance “trades” within the first $25 million of risk, due to the dynamics of the marketplace.

7.3. Aim of social policy

It would seem that the aim of social policy should be to support a free market for commercial liability insurance. Product liability tort reform (which would place con- straints on the amount of damages that can be awarded) and a two-tier regulatory climate (which permits federal chartering for insurance companies writing coverage for sophis- ticated commercial risks under federal regulation, limited to the solvency of the insurer only, while preserving state regulatory protection for individuals and small businesses) are but two initiatives that would support the flow of capital to the private insurance market.

The lasting lesson of Bhopal is that companies in hazardous industries should avail themselves of the high levels of coverage that are today attainable at reasonable prices and that the practice of purchasing commercial catastrophic insurance is socially desirable.

References

ACE Limited. (19931. ••Annual Report.”

Clements. R.. and R. Mendoza. (1994). -Risk is not a Font-Letter Word^^/>Vii/;r)/«/ (Marsh & McLentian Companies Quarterl)). Spring.

Kalelkar. A.S. (1988). ••ltnestigalion of Large-Magnitude Incidents: Bhopal as a Case Study.^’ Report presented at The Institution of Chemical Engineers Confetence on Preventing Major Chemical Accidents. May.

Redmond R.J. (1992). ••Stabilizing the Excess Liability Environment: How Scarcity was Turned into Capacity.” I’impoiiil (Marsh & McLennan Companies Quarterly). Fall.

Sinfield N. (1994).^^Asbestos—Human or Natural Disaster? Paper prepared for Conference on Social Treatment of Catastrophic Risk. Stanford Unixerslty. October.

Union Carbide Corporation. (19851. ••Bhopal Methyl Isocyanate Incident Investigation Team Report.” Danbury.

CT March.

Union Carbide Corporation. (1984). ••Annual Report.”